JPWO2004104845A1 - Storage system - Google Patents

Abstract

This is a storage system in which data is distributed to a plurality of storage devices and the redundancy of stored data is ensured. A storage device provided with a data storage unit composed of a plurality of sliced slice areas, a management information storage unit for storing management information of the slice areas, and communication having a communication function for data and information stored in the storage device It has multiple modules consisting of functional units, and these modules communicate with each other to copy the data stored in the slice area to an unused, arbitrary slice area of another module. Duplicate storage. Thereby, the safety of the data stored in the storage system is ensured.

Description

  The present invention relates to a storage system that stores data distributed to a plurality of storage devices, and more particularly to a storage system that ensures the redundancy of stored data.

Conventionally, in systems that process large volumes of information, when it is necessary to distribute and store data on multiple hard disks, etc., data can be read and written faster by accessing multiple hard disks in parallel. On the other hand, a storage system with redundancy is employed so that data is not lost even if one hard disk fails. One method for configuring this storage system is RAID.
RAID distributes data to multiple drives and strips simultaneously to improve access speed, writes the same data to multiple drives, mirroring to improve reliability over a single drive, data error Classification is made according to the level of fault tolerance, such as data restoration by detection. Today, RAID 0 has been devised from RAID 0 in which the access speed by striping is improved to RAID 5 in which parity can be added and data can be restored even if one of the dispersed disks fails. When a specific system is configured using these RAIDs, selection or combination is performed according to the level required for the system.
For example, although the appearance of RAID 1 is one virtual disk, it is actually composed of two physical disks, and each physical disk has completely the same contents, and is called mirroring. Therefore, when data is written to the virtual disk, data is simultaneously written to the two physical disks. Therefore, when the capacities of the two physical disks are different, it is necessary to match the physical disk having a smaller capacity. If a response indicating that the writing has been completed successfully is returned from both disks, an end response is sent to the data write request source, and if no response is returned from at least one of the disks, the data is returned. An error response is sent to the write request source. If one of the physical disks fails, data is read from the normal physical disk and written to the disk instead of the failed physical disk to maintain data redundancy.
For example, RAID 0 + 1 is a combination of RAID 0 and RAID 1. RAID 0 is a method in which a plurality of physical disks of the same capacity are divided into slices of the same size, a plurality of physical disks are ordered in advance, and slices are allocated in order, and is called data striping. Although this method does not ensure data redundancy, it has the characteristics that the access time can be reduced and the data read / write performance can be improved by accessing a plurality of physical disks simultaneously. On the other hand, RAID 0 + 1 is obtained by mirroring each slice of a physical disk, and has both RAID 0 characteristics and RAID 1 characteristics.
However, when a storage system is configured with RAID 0 + 1, a controller that executes various processes is required, as with RAID 1, and in order to handle stored data more safely, the controller is duplicated. In addition, since a communication path for accessing each physical disk from each of the duplicated controllers is required, the cost of the storage system becomes high. In addition, if one of the physical disks fails, even if there are many unused areas on the other physical disks, the unused area is assigned to the failed physical disk due to RAID 1 system restrictions. Since it cannot be used as an alternative, a spare physical disk is required.
On the other hand, the storage system is managed and controlled on the host computer side using the same interface for reading and writing data, and on the host computer independent of the interface for reading and writing data. There is. The former can be processed in cooperation with the host computer application and storage management control, but there is a problem that storage management control cannot be executed when the host computer is not activated. On the other hand, the latter can perform storage management control independently without depending on the platform of the host computer, but has the inconvenience that the application of the host computer and the storage management control cannot be processed in cooperation. There is.
Therefore, a system has been proposed in which a server for storage management control having an interface in both the host computer and the storage system is provided so that storage management control can be performed from the server (see Patent Document 1).
Further, a client accessing data stored in the storage system via a network cannot directly know how reliable the storage system is. For this reason, it is necessary to back up data in preparation for an unexpected situation, and for data that needs to be updated frequently, it is necessary to perform backup frequently in accordance with the update cycle, which is troublesome.
Therefore, the reliability level information is given to the data, and even if the reliability level is changed when the client refers to or updates the data, the data is changed according to the changed reliability level. There has been proposed a storage device having various reliability levels so that a storage device to be stored can be changed, which eliminates the need for a client to back up data (see Patent Document 2).
However, in the inventions disclosed in Patent Document 1 and Patent Document 2, the cost burden due to the duplication of the controller and the accompanying communication path, an unused area that cannot be used in a storage system configured with storage devices of various capacities, or It is not sufficient to solve problems such as the need for a spare physical disk.
SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a storage system that can increase the use efficiency of a storage device, reduce the duplication burden of a controller or the like, and ensure the safety of stored data. .
"Patent Document 1 JP 2002-268825 A (paragraph numbers 0012 to 0021, FIG. 1)"
“Patent Document 2 Japanese Laid-Open Patent Publication No. 2002-244922 (paragraph numbers 0012 to 0021, FIG. 1)”

The present invention provides a data storage unit having a plurality of sliced slice areas for storing data, a storage device having a management information storage unit for storing management information of the plurality of slice areas, and data stored in the data storage unit And a plurality of modules having a communication function unit having a communication function of information including management information stored in the management information storage unit, and stored by distributing and storing data in the data storage unit Data redundancy is ensured, the use efficiency of storage devices is increased, and the safety of stored data is increased at a low cost.
The plurality of modules can communicate with each other the information including the data stored in the data storage unit and the management information stored in the management information storage unit by the respective communication function units. In addition, the first module among the plurality of modules can copy the data stored in the slice area of the module to any slice area of the second module.
Therefore, by utilizing these communication functions and copying functions, this storage system stores the same data in each of the slice areas of at least two modules of the plurality of modules, and the slice in which the same data is stored. A segment composed of a pair of areas can be formed, thereby ensuring the redundancy of the data stored in the data storage unit. Furthermore, the redundancy of data makes it possible to restore the data by parity, and can improve fault tolerance.
The management information storage unit has management information such as unused slice regions and slice regions in which segments are formed in each data storage unit, as well as a slice region in which data that has received a storage request spans the slice regions of multiple modules. It is also possible to store information related to logical volumes expressed in units.
The management information stored in the management information storage unit is important in managing and controlling the storage system of the present invention. The management information storage unit of the first module among the plurality of modules is the second module, When an update request for the management information stored in the management information storage unit is received, the management information is updated when an update completion request is received from the second module, and when an update stop request is received, the management information Is restored to the state immediately before the management information update request is received, so that the current state can be reliably maintained.
In addition, when each of the plurality of modules is abnormally terminated, automatic recovery is not performed, but for example, a forced activation unit that performs manual activation is provided, and the forced activation unit can be restored by operating the operator.
Further, each of the plurality of modules is provided with a heartbeat sending unit that sends out a heartbeat (for example, “pulse signal arranged at a constant time interval”, the same applies hereinafter) that informs each operation state. It is possible to improve safety by periodically notifying each other of the operating state, recognizing each other's abnormality, and saving the data stored in the module in the abnormal state to another module.
Also, for example, an access module is provided that has information on data stored in each slice area of a plurality of modules, accepts access requests to the storage system from an external computer, and directly relays the access requests to the corresponding module. Thus, the access time can be shortened.
In addition, each of the modules is provided with a heartbeat sending unit that sends out a heartbeat that informs each operation state, and the heartbeats sent from the heartbeat sending unit are received to centralize the operation states of the modules. By providing a state management module that manages each of the modules, it is possible to reduce the load of each module and speed up measures when an abnormal state occurs in any of the modules.
In addition, management information is collected from each of the plurality of modules, and based on the collected management information, the data of the storage request is spread over the slice areas of the plurality of modules, and the information of the logical volume that represents the slice area as a unit is reproduced. By providing a volume information module to be built, speeding up operations by centralizing management information important for storage system operation, and storing data safety by duplicating the management information held by the storage information management section of each module It is possible to further improve the performance.
Furthermore, the storage system of the present invention is one of the above access module, state management module, and volume information module depending on the nature and load of the information processing system that accesses the storage system, and the data reliability level. Alternatively, all can be used in combination with a module having the data storage unit and the management information storage unit.
With the above configuration of the present invention, it is possible to provide a storage system in which the use efficiency of storage devices is high, the safety of stored data is ensured at low cost.

FIG. 1 is a diagram illustrating a configuration example of an information processing system including the storage system according to the first embodiment.
FIG. 2 is a diagram for explaining the basic concept of storing data in four storage devices in the storage system of this embodiment.
FIG. 3 is a functional block diagram illustrating the storage system according to the first embodiment.
FIG. 4 is a diagram showing an example of a processing procedure when copying data stored in the slice area in the storage system.
FIG. 5 is a diagram showing an example of a processing procedure when there is a data storage request from the outside to the storage system.
FIG. 6 is a diagram illustrating an example of a processing procedure when there is a data storage request from the outside to the storage system.
FIG. 7 is a diagram showing an example of a processing procedure when there is a data storage request from the outside to the storage system.
FIG. 8 is a diagram showing an example of a processing procedure when there is a data storage request from the outside to the storage system.
FIG. 9 is a diagram illustrating an example of a processing procedure when there is a data storage request from the outside to the storage system.
FIG. 10 is a diagram illustrating an example of a processing procedure when there is a data storage request from the outside to the storage system.
FIG. 11 is a diagram illustrating an operation when a release request is issued while the storage system is operating.
FIG. 12 is a diagram illustrating an operation when a release request is issued while the storage system is operating.
FIG. 13 is a diagram illustrating an operation when a release request is issued while the storage system is operating.
FIG. 14 is a diagram illustrating an example of a processing procedure when any module of the storage system fails.
FIG. 15 is a diagram illustrating an example of a processing procedure when any module of the storage system fails.
FIG. 16 is a diagram illustrating an example of a processing procedure when any module in the storage system fails.
FIG. 17 is a diagram illustrating an example of a processing procedure for extracting any module of the storage system.
FIG. 18 is a diagram showing an example of a processing procedure for extracting any module of the storage system.
FIG. 19 is a diagram illustrating an example of a processing procedure for extracting any module of the storage system.
FIG. 20 is a diagram illustrating an example of a processing procedure for replacing any module in the storage system with a new module.
FIG. 21 is a diagram illustrating an example of a processing procedure for replacing any module in the storage system with a new module.
FIG. 22 is a diagram illustrating an example of a processing procedure when replacing any module of the storage system with a new module.
FIG. 23 is a diagram illustrating an example of a processing procedure when data stored in a module and management information are updated.
FIG. 24 is a diagram illustrating an example of a processing procedure when data stored in a module and management information are updated.
FIG. 25 is a diagram illustrating an example of a processing procedure when data stored in a module and management information are updated.
FIG. 26 is a diagram illustrating the module A11 as a representative example.
FIG. 27 is a diagram illustrating the module A11 as a representative example.
FIG. 28 is a diagram illustrating a configuration example of an information processing system including the storage system according to the second embodiment.
FIG. 29 is a functional block diagram showing the storage system of the second embodiment.
FIG. 30 is a diagram illustrating a configuration example of an information processing system including the storage system according to the third embodiment.
FIG. 31 is a functional block diagram showing the storage system of the third embodiment.
FIG. 32 is a diagram illustrating an example of functions and processing procedures of the state management module N of the present embodiment.
FIG. 33 is a diagram illustrating an example of functions and processing procedures of the state management module N of the present embodiment.
FIG. 34 is a diagram illustrating another example of functions and processing procedures of the state management module N of the present embodiment.
FIG. 35 is a diagram showing a case where the state management module N of the present embodiment shown in FIG. 34 has a function for forcibly terminating the operation of a module in which an abnormality has occurred.
FIG. 36 is a diagram illustrating a case where the state management module N of the present embodiment illustrated in FIG. 35 has a function of transmitting state information to each module.
FIG. 37 is a diagram illustrating a configuration example of the volume information module M of the present embodiment.
FIG. 38 is a diagram showing a data structure of module information in the module information section.
FIG. 39 is a diagram showing a data structure of logical volume information in the logical volume unit.
FIG. 40 is a diagram showing an example of the management information update function that the volume information module M has.
FIG. 41 is a diagram showing an example of the management information update function that the volume information module M has.
FIG. 42 is a diagram illustrating an example of a copy request function for data stored in a module included in the volume information module M.
FIG. 43 is a diagram illustrating an example of a function of the volume information module M that forcibly terminates processing in each module.
FIG. 44 is a diagram showing an example of a processing procedure when the volume information module M receives a data storage request.
FIG. 45 is a diagram showing an example of a processing procedure when the volume information module M receives a data storage request.
FIG. 46 is a diagram illustrating an example of a processing procedure when the volume information module M receives a data storage request.
FIG. 47 is a diagram showing an example of a processing procedure when the volume information module M receives a logical volume release request.
FIG. 48 is a diagram showing an example of a processing procedure when the volume information module M receives a logical volume release request.
FIG. 49 is a diagram showing an example of a processing procedure when the volume information module M receives a logical volume release request.
FIG. 50 is a diagram illustrating an example of a processing procedure when a predetermined module fails.
FIG. 51 is a diagram illustrating an example of a processing procedure when a predetermined module fails.
FIG. 52 is a diagram illustrating an example of a processing procedure when a predetermined module fails.
FIG. 53 is a diagram illustrating an example of a processing procedure when a predetermined module fails.
FIG. 54 is a diagram illustrating an example of a processing procedure for extracting a predetermined module.
FIG. 55 is a diagram illustrating an example of a processing procedure for extracting a predetermined module.
FIG. 56 is a diagram illustrating an example of a processing procedure when a predetermined module is pulled out.
FIG. 57 is a diagram showing an example of a processing procedure when modules are exchanged.
FIG. 58 is a diagram showing an example of a processing procedure when modules are exchanged.
FIG. 59 is a diagram illustrating an example of a processing procedure when modules are replaced.
FIG. 60 is a diagram illustrating an example of a processing procedure when modules are exchanged.
FIG. 61 is a diagram illustrating a configuration example of an information processing system including the storage system according to the fourth embodiment.
FIG. 62 is a functional block diagram showing the storage system of the fourth embodiment.

(First embodiment)
FIG. 1 is a diagram showing a configuration example of an information processing system including a storage system according to the first embodiment of this invention.
In FIG. 1, the storage system 1 includes a module A <b> 11 including a server having a storage device 21 (“personal computer or workstation”; the same applies hereinafter), a module B <b> 12 including a server having a storage device 22, and a server having a storage device 23. And a module D14 including a server having a storage device 24. Each of these four modules is connected to the network 2 via the communication line 4. Also, two computers 3 are connected to the network 2, and each of the two computers 3 accesses the storage system 1 via the network 2 and makes a request for storage, addition, deletion, change, etc. of data. It is configured to be able to. Each of the storage devices 21 to 24 is configured by a hard disk here, but is not limited thereto.
The four modules constituting the storage system 1, that is, the module A 11, the module B 12, the module C 13, and the module D 14, are controlled by the respective servers, and not only between the network 2 but the network 2 due to the communication function of the servers. In addition, it has a configuration capable of sending and receiving information including data between modules connected via a communication line. Therefore, by copying the data stored in the storage devices 21 to 24 of the modules 11 to 14, the same data can be stored in duplicate in the storage devices with different modules. The redundancy of the data stored in can be ensured.
FIG. 2 is a diagram for explaining the basic concept of storing data in four storage devices in the storage system of this embodiment.
As shown in FIG. 2, each of the four storage devices 21 to 24 constituting each module of the present embodiment does not necessarily have the same capacity, but a plurality of small areas S (numbers S11) that store data of a certain capacity. To S15, numbers S21 to S26, numbers S31 to S35, and numbers S41 to S45), and the small areas S are referred to as slice areas 15, respectively. When a data storage request is issued to the storage system, the data is stored in units of slice areas 15. A group of data for which a storage request has been received, across slice areas of a plurality of modules, is referred to as a logical volume.
For example, in the configuration shown in FIG. 2, one logical volume is configured by slice areas S11, S21, and S23 indicated by oblique diagonal lines from the upper right.
Further, in the storage system of this embodiment, different modules in units of each slice area 15 or the same data as the data stored in the same module is stored in any module different from the stored slice area. , And stored in any one of the unused slice areas 15.
Thus, each storage device can be used efficiently while ensuring the redundancy of the stored data.
A pair of slice areas 15 in which the same data is stored is referred to as a segment 17. This segment 17 can be constructed by a module that has received a data storage request by copying the data stored in the slice area 15 to an unused slice area 15 of another module using the communication function. it can.
For example, the slice region S11 indicated by the oblique diagonal line from the upper right constitutes the segment 17 and the slice region S22 indicated by the oblique oblique line from the upper left, and the slice region S21 indicated by the oblique oblique line from the upper right is indicated by the oblique oblique line from the upper left. A slice region S23 that constitutes the slice region S42 and the segment 17 and is indicated by the oblique diagonal line from the upper right constitutes a slice region S32 and the segment 17 that is indicated by the oblique oblique line from the upper left.
Here, although the number of modules is four in this embodiment, it is not limited to this number. Each module is configured to be connected to the network via the communication line 4, but may be directly connected to the network. Furthermore, by forming a segment in units of slice areas and duplicating data, the redundancy of the data stored in the storage device is ensured, but data error detection by parity check and data error detection Redundancy of stored data may be ensured by using it together with correction.
FIG. 3 is a functional block diagram illustrating the storage system according to the first embodiment.
The storage system illustrated in FIG. 3 includes a module A11, a module B12, a module C13, and a module D14. Each of the modules 11 to 14 has a network 2 connected to the server by communication function units 25 to 28 included in the server. The data can be transmitted / received to / from the outside. The servers are also connected by a one-to-one communication line 4, and communication can be performed between modules by the communication function units 25 to 28. Here, the servers are connected by a one-to-one communication line 4, but it is not always necessary to connect the servers by a one-to-one communication line 4, and any connection form can be used as long as communication is possible between modules. be able to.
The storage devices 21 to 24 of each of the module A11, the module B12, the module C13, and the module D14 store the data storage unit 8 having a plurality of slice areas 15 for storing data and management information for the plurality of slice areas 15. Management information storage unit 18. The management information stored in each management information storage unit 18 includes, for example, each slice area (S11 to S15, S21 to S26, S31 to S35, S41 to S45), a logical volume name Ln, and an arrangement position of data on the logical volume. n and slice regions of other modules belonging to the same segment are arranged in this order.
In the storage system of this embodiment, there are two logical volumes named L1 and L2. The logical volume L1 has three segments, the data is stored in the slice areas S11, S23, and S43 in the order of the data arrangement positions, and the segment is formed by the slice areas S11 and S22, S23 and S32, and S43 and S24. Is configured. The logical volume L2 has two segments and is stored in the slice areas S14 and S43 in the order of the data arrangement positions, and the slice areas S14 and S33 and S43 and S24 constitute a segment.
Here, the slice area 15 and the management information storage unit 18 are arranged in different areas on the storage device, but the arranged areas can also be mixed.
As described above, since data is distributed and stored for each slice area 15 divided into small capacities by slicing, even if the capacities of the storage devices of each module are different from each other, the respective storage devices Can improve the efficiency of use. In addition, since the module has a function of communicating with each other and has a function of copying data stored in each module to another module, redundancy is ensured by duplication of stored data.
Hereinafter, an operation when data is stored in the storage system will be described.
FIG. 4 is a diagram showing an example of a processing procedure when copying data stored in the slice area in the storage system.
In FIG. 4, the storage devices 21 to 24 of the respective modules 11 to 14 include a slice area 15 for storing data requested to be written, the logical volume name Ln to which the stored data belongs, and the data on the logical volume. Management information storage for storing the management information recorded for each slice area (S11 to S15, S21 to S26, S31 to S35, S41 to S45) of the slice area 15 of another module belonging to the same segment as the arrangement position n A portion 18 is provided.
Now, for some reason, when it becomes necessary to copy the data stored in the slice area S14 of the module A11 to the slice area S41 of the module D14, the module A11 activates the communication function unit 25, The channel with the communication function unit of the module 14 is connected, and an inquiry is made as to whether or not S41 is unused. When receiving a reply from module D14 indicating that S41 is not used, module A11 transmits the data stored in number S14 to module 14.
FIG. 5 to FIG. 10 are diagrams illustrating an example of a processing procedure when there is a data storage request from the outside to the storage system.
The storage system shown in FIGS. 5 to 10 includes a module A11, a module B12, a module C13, and a module D14. Each module has communication function units 25 to 28, and the storage devices 21 to 24 have a plurality of slices. Although the data storage unit 8 including the area 15 and the management information storage unit 18 are provided, the display of the data storage unit 8 is omitted for convenience.
In FIG. 5, the module A11 stores data in slice area numbers S11 and S14, and the slice areas S12, S13, and S15 are unused. In the module B12, data is stored in the slice areas S21, S22, S23, and S24, and the slice areas S25 and S26 are unused. The module C13 stores data in the slice areas S32 and S33, the slice areas S31 and S34 are unused, and the module D14 stores data in the slice areas S41 and S42, and the slice areas S41, S44, and S45. Is unused.
Here, a case will be described as an example where the module A11 has a storage request for a logical volume with four segments.
Various rules can be considered for logical volume allocation. Here, logical volumes are allocated under the rule of preferentially storing them in their own slice area as much as possible. Each module requests a different module to configure a segment by securing an empty slice area and copying data.
As shown in FIG. 6, since the slice areas S12, S13, and S15 are unused, the module A11 allocates data of the logical volume name L3 to these slice areas. Next, in order to search for slice areas of other modules constituting the slice area and the segment, first, the module B12 is inquired about an unused slice area and a request for securing an unused slice area.
As shown in FIG. 7, since the module B12 has two unused slice areas, slice areas S25 and S26, it returns to the module A11 that they are allocated as the segment 1 and the segment 2 of the logical module name L3. Module A11 copies the data stored in number S12 to slice area S25, and copies the data stored in slice area S13 to slice area S26. A segment is configured by the slice areas S12 and S25 and the slice areas S13 and S26.
Since the module A11 does not have enough slice areas that constitute a segment, only the module A11 makes a similar inquiry and request to the module C13 as shown in FIG. The module C13 does not use the slice areas S31 and S34, but only one slice area is requested from the module A11, so only the slice area S31 is secured. Then, as illustrated in FIG. 9, the module C13 returns to the module A11 that the segment 3 is allocated to the slice area S31. The module A11 copies the data stored in the slice area S15 to the slice area S31, and the slice areas S15 and S31 constitute a segment.
Next, since the module A11 could not secure the slice area for the number of segments requested for allocation by itself, the module A11 requests the module B12 to allocate a logical volume. However, since there is no longer any unused slice area in module B12, the logical volume allocation request moves to module C13.
Since the module C13 has an unused slice area S34, it is assigned as the logical volume name L3. Then, the module D14 is inquired about an unused slice area and requests to secure an unused slice area to construct a segment. Since module D14 has three unused slice areas, it returns to module C13 that slice area S41, which is one of them, is segment 4.
The module C13 copies the data stored in the slice area S34 to the slice area S41, and configures a segment by the slice areas S34 and S41 as shown in FIG. As a result, the logical volume of the requested number of segments is secured, and the module D14 returns the completion of allocation to the request source.
Here, the return to the request source is not necessarily performed by the module D14, and may be returned to the request source after the module A11 receives the completion notification from the module D14.
By taking such a procedure, redundancy is ensured by duplication of stored data while increasing the storage device use efficiency of each module.
Although the case where a logical volume of the requested size could be secured has been described here, if allocation of a logical volume fails due to insufficient slice area to be allocated, at that stage The module that has received the allocation request, the module that has received the inquiry, or the like can return an allocation stop request to the request source. As a result, the slice area that has been secured halfway can be released to correspond to another logical volume. Although the storage request has been described here, the addition, deletion, and change requests can be performed in the same manner.
Next, a case where stored data is moved in module units will be described.
FIG. 11 to FIG. 13 are diagrams showing the operation when a release request is issued while the storage system is operating.
The storage system shown in FIGS. 11 to 13 includes a module A11, a module B12, a module C13, and a module D14. Each of the modules 11 to 14 includes communication function units 25 to 28. The storage devices 21 to 24 include Although a data storage unit 8 including a plurality of slice areas 15 and a management information storage unit 18 are provided, the display of the data storage unit is omitted for convenience.
In FIG. 11, the module B12 receives a release request for the logical volume name L2. Then, the module B12 notifies the release instruction of the logical volume name L2 to the other modules A11, C13, and D14 as shown in FIG. Each module that has received the notification checks the number of the slice area to which the logical volume name L2 is assigned from the information stored in the management information storage unit 18 in its own module. Then, as shown in FIG. 13, the data is deleted from the allocated slice areas S14, S24, S33, 43, and the information in the management information storage unit 18 is also deleted, so that the module A11, the module C13, and the module D14 The release completion is notified to the module B12, and the module B12 returns a request completion response to the request source.
Next, FIG. 14 to FIG. 16 are diagrams showing an example of a processing procedure when any module of the storage system fails.
The storage system shown in FIGS. 14 to 16 includes a module A11, a module B12, a module C13, and a module D14. Each of the modules 11 to 14 has communication function units 25 to 28, and the storage devices 21 to 24 include Although a data storage unit 8 including a plurality of slice areas 15 and a management information storage unit 18 are provided, the display of the data storage unit 8 is omitted for convenience.
In FIG. 14, when the module B12 fails, the other modules A11, C13, and D14 recognize the failure by some means. The other modules A11, C13, and D14 are based on the information stored in the management information storage unit 18, and the slice areas S11, S32, S42, and the slice areas that belong to the module B12 and the segments. S44 is detected. Then, since it is necessary to secure a slice area that newly constitutes a segment with each of these slice areas, as shown in FIG. 15, module A11 is from module C13, module 13 is from module D14, and module D14 is , An unused slice area is secured from each module A11. The slice area S11 constitutes a segment with the slice area S31, the slice area S32 constitutes a segment with the slice area S44, the slice area number S42 becomes the slice area S12, and the slice area number S43 becomes the slice area S13. Configure the segment.
Next, as shown in FIG. 16, the module A11, the module C13, and the module D14 copy the data stored in the respective slice areas S11, S32, S42, and S44 to the counterpart module that constitutes the segment. . When copying of each data is completed, the management information stored in each management information storage unit is determined.
By taking such a procedure, it is possible to respond to a predetermined module release request without losing the redundancy of the data stored in each module.
Next, FIG. 17 to FIG. 19 are diagrams showing an example of a processing procedure in the case of extracting any module of the storage system.
The storage system illustrated in FIGS. 17 to 19 includes a module A11, a module B12, a module C13, and a module D14. Each of the modules 11 to 14 includes communication function units 25 to 28, and the storage devices 21 to 24 include Although a data storage unit 8 including a plurality of slice areas 15 and a management information storage unit 18 are provided, the display of the data storage unit 8 is omitted for convenience.
As shown in FIG. 17, when the module D14 receives the extraction request, the module D14 is a destination of data stored in the slice areas S42 and S43 of the module D14, and configures a segment with each of the slice areas S42 and S43. It is necessary to search for the slice areas S21 and S24 and a slice area newly constituting a segment. Therefore, the module B12 to which the numbers S21 and S24 constituting the segments and the numbers S42 and S43 of the slice area of the module D14 belong, respectively, is not suitable as a data movement destination.
As shown in FIG. 18, the data stored in the numbers S42 and S43 are moved to the modules A11 and C31, respectively, and the segments are respectively formed in the free numbers S31 and S12 and the numbers S42 and S43. The data stored in the slice areas S42 and S43 is copied to the slice areas S21 and S24.
When copying is completed, as shown in FIG. 19, module D14 notifies module A11, module C13, and module B12 that the segment configuration has been changed, and module D14 can be pulled out. Note that the management information storage unit 18 of the module A11, the module B12, and the module C13 changes the stored management information.
By taking such a procedure, it is possible to respond to the release request of the module without losing the redundancy of the data stored in the module.
Next, FIGS. 20 to 22 are diagrams illustrating an example of a processing procedure in the case where any module of the storage system is replaced with a new module.
Here, after the module D14 is pulled out, a new module E55 is added, and data stored in the module C13 is moved to the module E55. The action of pulling out the module D14 is already the same as the action described with reference to FIGS.
In the storage systems shown in FIGS. 20 to 22, the module D14 out of the module A11, the module B12, the module C13, and the module D14 is pulled out and a new module E55 is added. Each of the modules 11 to 13 and 55 includes communication function units 25 to 28, and the storage devices 21 to 23 and 240 include a data storage unit 8 including a plurality of slice areas 15 and a management information storage unit 18. However, the display of the data storage unit 8 is omitted for convenience.
As illustrated in FIG. 20, it is assumed that the addition work of the new module E55 is completed, and for example, the module C13 receives an exchange command for the module E55.
As illustrated in FIG. 21, the module C13 that has received the exchange command confirms whether the three slice areas for moving data can be secured in the module E55. When it is confirmed that the slice area can be secured, as shown in FIG. 22, the data stored in the slice areas S31, 332, and S33 are copied to the slice areas S51, S52, and S53, respectively. When the copying is completed, the information stored in the management information storage unit 18 of the modules A11 and B12 to which the slice areas constituting the segments and the slice areas S31, 332, and S33 belong is updated. As a result, all the data stored in the module C13 has moved to the module E55, so that the module C13 is pulled out and the exchange is completed.
By taking such a procedure, modules can be exchanged without losing redundancy of stored data.
Next, updating of management information stored in the management information storage unit will be described.
23 to 25 are diagrams illustrating an example of a processing procedure when the management information stored in the module is updated.
23 to 25, the module A11 and the module B12 of the storage system are shown as representatives in order to simplify the description.
The modules are connected to each other via a one-to-one communication line, and can communicate with each other via the communication function units 25 and 26. The storage devices 21 and 22 of the modules A11 and B12 respectively have a data storage unit 8 including a plurality of slice areas 15 for storing data requested to be written, and a management information storage unit 18, and store management information The unit 18 includes a logical volume name Ln to which the stored data belongs, an arrangement position n of data on the logical volume, and a slice area 15 of another module belonging to the same segment, in each slice area (S11 to S15, S21 to S21). In addition to the management information recorded every S26), update processing information indicating that the stored management information is being updated (Δ mark) and has been updated (◯ mark) is added. In addition, a pre-processing information storage area 19 is provided for temporarily storing management information before entering the update process in order to cope with an unexpected situation during the update process of the management information.
Here, since the slice areas S15 and S25 of the module A11 and the module B12 are being updated, a Δ mark is given.
The pre-processing information storage area 19 may be mixed with the slice area 15 or the management information storage unit 18 of each storage device 21 or 22, or an independent area may be provided.
When the module A11 and the module B12 receive the update completion request, as shown in FIG. 24, the management information is confirmed and the update processing information of the numbers S15 and S25 in the management information storage unit 18 has been updated (circle mark). Change to
On the other hand, when the module A11 and the module B12 receive an update stop request during the update process, as shown in FIG. 25, the management information temporarily stored in the pre-process information storage area 19 is copied, The state before the update is restored, and the update processing information in the slice areas S15 and S25 of the management information storage unit 18 is changed to updated (◯ mark).
Here, a case where one update process is performed has been described, but a plurality of update processes can be simultaneously performed in parallel by using an identifier instead of a circle mark or a triangle mark.
Next, transfer of status information of each module will be described.
26 and 27 are diagrams showing the module A11 as a representative example.
The module A11 shown in FIG. 26 includes the communication function unit 25 and the storage device 21, and also includes a switch 30 that forcibly activates the module. The storage device 21 includes a data storage unit 8 including a plurality of slice regions 15 and a management information storage unit 18, and also includes a state information storage region 20 for storing state information when the module A11 finishes operation. Have. In the status information storage area 20, “normal information” is stored as a storage device when the operation ends normally, and “abnormal information” is stored as a storage device when the operation ends due to an unexpected accident such as a power failure during operation. The module A11 is configured to determine whether to start up with reference to information stored in the state information storage area 20. In this case, since the module can be activated by operating the switch 30, even if an unexpected accident occurs, it can be dealt with.
In addition to the module A11 shown in FIG. 26, the module A11 has a heartbeat sending unit 31 that sends the status information of the module A11 to other modules via the communication function unit 25 in the storage device 21. . This heartbeat sending unit 31 is informed that if there is no problem in the module when the module is activated, it is in a normal state, and informs that it is abnormal if an abnormal state occurs during module activation and normal operation. Fulfill. Therefore, other modules that receive the notification from the module A11 can achieve early recovery of stored data and information.
Here, the heartbeat transmission unit 31 may periodically transmit the state information. In this way, if status information is sent periodically, it is possible to notify other modules that the normal status continues, and if the status information is not received for a certain period of time, there is an error in that module. Since it can be considered that a condition has occurred, a quick response can be made. Further, when communication with another module becomes impossible due to a failure of its own communication function unit 25 or a failure of the communication line 4, the safety of the entire data stored in the storage system is obtained by performing its own termination process. Can also be increased.
(Second Embodiment)
The storage system of this embodiment is different from the storage system of the first embodiment in that it includes an access module that directly accesses the corresponding slice area when an access request to the storage system is received. Other points are common. Therefore, the different access modules will be mainly described below.
FIG. 28 is a diagram illustrating a configuration example of an information processing system including the storage system according to the second embodiment.
The storage system 1 shown in FIG. 28 includes a module A11 including a server (personal computer or workstation; the same applies hereinafter) having a storage device 21, a module B12 including a server having a storage device 22, and a server having a storage device 23. When receiving an access request to the module C13, the server having the storage device 24, and the storage system, a personal computer (hereinafter abbreviated as "PC") having information that can directly access the slice area of the corresponding module. Two access modules A41 and A42. Each of the four modules is connected to the access module via the communication line 4, and the access module is connected to the network 2. Therefore, each of the two computers 3 connected to the network 2 is configured to first access one of the access modules and access each module via the access module.
FIG. 29 is a functional block diagram showing the storage system of the second embodiment.
The storage system of this embodiment is common to the storage system of the first embodiment described with reference to FIG. 3 except that it includes access modules A41 and A42 made up of PCs. Therefore, the same components are denoted by the same reference numerals, and redundant description is omitted.
The storage system shown in FIG. 29 has a module A11, a module B12, a module C13, and a module D14. Each of the modules 11 to 14 has two access modules made up of PCs by communication function units 25 to 28 of the server. Connected to both sides. The modules are connected to each other by a one-to-one communication line 4, and the modules 11 to 14 can communicate with each other by the communication function units 25 to 28. An external computer connected to the network 2 can directly access a predetermined module by accessing one of the two access modules.
Each of the storage devices 21 to 24 of the module A11, the module B12, the module C13, and the module D14 has a data storage unit 8 including a plurality of slice areas 15 that store data requested to be written, and a logical to which the stored data belongs. For each slice area (S11 to S15, S21 to S26, S31 to S35, S41 to S45), the volume name Ln, the data arrangement position n on the logical volume, and the slice area 15 of another module belonging to the same segment And a management information storage unit 18 for storing the recorded management information.
Each access module contains information such as the logical volume name, the module and slice area in which data at each position of the logical volume is stored, the slice area that constitutes the slice area and segment, and the unused slice area in each module. A table consisting of Therefore, when there is a request for addition, change, or deletion of data stored in a slice area or a request to store data in an unused slice area 15, a module corresponding to the request is searched for and accessed. Requests can be relayed directly to the corresponding module. In this case, the access module may have a table including all information stored in the storage system, or a part of the information is stored and information that satisfies an external request is not stored. In this case, a new table may be constructed by deleting unnecessary information, inquiring each module for information that is not held, and rewriting information held as a table.
In this way, an access module having a table is provided, the access module receives an access request from the outside, and the access module directly accesses a predetermined module by referring to the table. Access time can be shortened. Further, the two access modules are operated independently here, but the security can be enhanced by, for example, a duplex configuration.
(Third embodiment)
Compared to the first embodiment, the third embodiment of the present invention is a storage system in which a status management module that manages the status of each module and a logical volume of data that is distributed and stored in each module. A volume information module for reconstructing information, and the operation of the storage system is led by the volume information module. However, since the other points are common, the differences will be mainly described below.
FIG. 30 is a diagram illustrating a configuration example of an information processing system including the storage system according to the third embodiment.
In FIG. 30, the storage system 1 includes four modules G31, a module H32, a module I33, and a module J34 that are PCs having storage devices, and each of the four modules is connected to the network 2 via the communication line 4. Has been. The PC 2 manages the state of each module, reconstructs information for each logical volume of data distributed and stored in each module, and issues a command to each module based on the information. Two management control modules 61 and 62 are provided, and the two management control modules 61 and 62 are respectively connected to the network 2.
The two management control modules 61 and 62 have both functions of the state management module of the present invention and the volume information module, and one of the two management control modules 61 and 62 checks the operation of the other, Even if trouble occurs on one side, a dual system in which processing is continued by the other may be configured, or a duplex system may be configured to operate simultaneously and use one as the main and the other as an emergency backup. By taking such a configuration of the management control module, the safety of the management control function of the storage system can be enhanced.
Here, in the present embodiment, the functions of both the state management module and the volume information module are integrated as a management control module. However, such a configuration is not necessarily required, and each can be configured as an independent module. .
Hereinafter, for convenience of explanation, the state management module N and the volume information module M will be described based on functional blocks configured as independent modules.
FIG. 31 is a functional block diagram showing the storage system of the third embodiment.
The storage system shown in FIG. 31 includes a module G31, a module H32, a module I33, and a module J34. Each of the modules 31 to 34 has the network 2 connected to the PC by the communication function units 25 to 28 of the PC. The data can be transmitted to and received from an external computer. The PCs are also connected by a one-to-one communication line 4, and the modules 31 to 34 can communicate with each other by the communication function units 25 to 28 of the PC. Here, the PCs are connected by a one-to-one communication line 4, but it is not always necessary to connect them by a one-to-one communication line 4, and any topology can be used as long as communication is possible between modules. Can do.
In each of the storage devices of the module G31, the module H32, the module I33, and the module J34, a data storage unit 8 including a plurality of slice areas 15 for storing data requested to be stored, and a logical volume name Ln to which the stored data belongs. , Management position recorded on the logical volume, and the slice area 15 of another module belonging to the same segment, recorded for each slice area (S11 to S15, S21 to S26, S31 to S35, S41 to S45) A management information storage unit 18 for storing information is provided, but the display of the data storage unit 8 is omitted here.
In addition, the storage system according to the present embodiment collects management information from each module and monitors the status of each module, collects management information from each module, and monitors the status of each module. Or a volume information module M64 for centrally grasping information for each logical volume.
Each of the state management module N63 and the volume information module M64 has a communication function unit 29, and each is connected to the network 2 and configured to be able to communicate with each of the modules 31 to 34 via the network 2. Yes. Further, the state management module N63 and the volume information module M64 can communicate with each other. Therefore, the volume information module M64 requests the state management module N63 to end the processing of a specific module, for example, or ends the operation of the entire storage system according to the information of each module received from the state management module N63. It can also be made.
Next, details of the functions of the state management module N63 and the volume information module M64 and their operations will be described.
32 and 33 are diagrams illustrating an example of the function and processing procedure of the state management module N of the present embodiment.
32, the storage system includes a module G11, a module H12, a module I13, and a state management module N.
Here, the state management module N is installed independently of each module, but the function may be shared by the PC of any module.
Each module sends to each storage device status information of each module (for example, a heartbeat consisting of a pulse signal at regular intervals. When the heartbeat is received, it indicates a normal state, and when it is not received, it is abnormal. A heartbeat sending unit 21a, 22a, 23a, and 24a that transmits the state) to other modules and the state management module via the communication function unit. The heartbeat sending units 21a, 22a, 23a, and 24a inform that the module is in a normal state when there is no problem in the module when the module is started, and abnormal if an abnormal state occurs during module startup and normal operation. Can be informed. Therefore, other modules that receive notifications from any of the heartbeat sending units can achieve early recovery of stored data and information.
The state management module N63 includes a heartbeat receiving unit 63a that receives state information sent from the heartbeat sending units 21a, 22a, 23a, and 24a of the modules 31 to 34, and a module state management table that manages the received state information. And. In the module state management table 63b, whether each module is normal or abnormal is displayed by, for example, a circle mark and a cross mark.
The module G31, the module H32, and the module I33 transmit the respective state information to the state management module N63. The state management module N63 receives the state information sent from each module and displays it on the module state management table 63b.
Here, even if the information displayed in the module state management table 63b is lost, it can be acquired from each module again. Therefore, it is not always necessary to set the information in the secondary storage device of the PC. May be.
Next, as shown in FIG. 33, for example, when the module I33 detects an abnormal state in its own module and notifies the state management module N63 of the abnormal state, the state management module N63 performs module state management related to the module I33. The display in Table 63b is changed to a cross.
FIG. 34 is a diagram illustrating another example of functions and processing procedures of the state management module N of the present embodiment.
As shown in FIG. 34, each module is configured to send status information (heartbeat) at regular time intervals. Further, the state management module N includes a time monitoring unit 63c that records the time of the state information that has been sent, and the time when the latest state information is received is displayed in the module state management table 63b.
Since the status information is not transmitted from the module I for a certain period of time, the status management module N63 determines the status of the module I33 from the circle mark by the time monitoring unit 63c. Change to mark. this? The mark represents that it is estimated that an abnormality has occurred because no state information is sent.
FIG. 35 is a diagram showing a case where the state management module N of the present embodiment shown in FIG. 34 has a function for forcibly terminating the operation of a module in which an abnormality has occurred.
As shown in FIG. 35, the state management module N63 has a forced termination command sending unit 63d that sends a forced termination command to the corresponding module, and each of the modules 31 to 34 is sent from the forced termination command sending unit 63d. The forced termination command receivers 21b, 22b, 23b, and 24b are provided with a function of receiving the forced termination command and terminating the processing being executed.
Here, the state management module N63 may also include a forced termination command receiving unit.
In this way, by providing the state management module N63 with a forced termination command sending function, when an abnormality in each module is recognized, the processing in that module is immediately terminated and temporarily removed from the storage system. The processing can be continued by a simple module.
FIG. 36 is a diagram illustrating a case where the state management module N of the present embodiment illustrated in FIG. 35 has a function of transmitting state information to each module.
As illustrated in FIG. 36, the state management module N63 includes a module state transmission unit 63e that transmits state information grasped by the state management module N63 to each of the modules 31 to 34. Module state receivers 21c, 22c, 23c, and 24c that receive state information transmitted from the state transmitter 63e are included.
Here, the module state transmitting unit 63e may transmit the state information only to the module that is recognized as having an abnormality, or transmits the state information to all modules including normal ones. You may decide. In that case, each module receives only relevant state information.
As a result of comparing the state information transmitted from the module state transmission unit 63e with the state information held by each module 31 to 34, it is determined that there is a contradiction or that it is preferable to terminate the processing of itself. If so, the process is terminated.
In this way, if the state information held by the state management module N is compared with the state information held by the state management module N and priority is given to the information managed by the state management module N, storage based on consistent information is performed. The entire system can be operated stably.
FIG. 37 is a diagram illustrating a configuration example of the volume information module M of the present embodiment.
37, the storage system includes a module G31, a module H32, a module I33, a module J34, and a volume information module M64. These modules 31 to 34 and 64 have communication function units 25 to 29, respectively.
The volume information module M64 has a communication path for the network 2 and also has a communication path for each of the module G31, the module H32, the module I33, and the module J34. Further, the module G31, the module H32, the module I33, and the module J34 are connected to each other via the communication line 4.
Each of the modules 31 to 34 includes a data storage unit 8 including a plurality of slice areas 15 and a management information storage unit 18 in the storage devices 21 to 24, and stores the management information stored in the management information storage unit 18 as a volume information module. Although it has the function to transmit to M64, the display of the data storage part 8 is abbreviate | omitted for convenience.
The volume information module M64 includes a management information collection unit 64a that collects management information stored in the management information storage unit 18 from each of the modules 31 to 34 via the communication function unit 29, and a module status from the collected management information. And a logical volume unit 64c for reconstructing information for each logical volume indicating the slice area 15 in which each position of the logical volume is stored.
Here, the module information and the logical volume information reconstructed by the volume information module M64 are constructed based on the management information of the modules 31 to 34, and can be reconstructed even if they disappear. What is necessary is just to hold | maintain in a main storage device, and it does not necessarily need to preserve | save in a secondary storage device.
FIG. 38 is a diagram showing a data structure of module information in the module information section.
As shown in FIG. 38, the module information includes a first file 110 for managing the module state and a second file 120 for managing an unused slice area. The first file and the second file are Both are linked to the management information file 130 of each module. The management information file 130 has a pointer 131 to the next management information and information 133 for each slice area. The first file 110 includes a pointer 111 to the next management information, a module name 112, a module state 113, a pointer 114 to a slice area of management information held by the module, and a pointer 115 to an unused slice area. And has as many entries 100 as the number of modules capable of communicating with the volume information module M64. The second file 120 has a pointer 121 to the next management information, a pointer 122 to the next unused slice area, and information 123 for each slice area.
FIG. 39 is a diagram showing a data structure of logical volume information in the logical volume unit.
As shown in FIG. 39, the logical volume information includes a third file 210 that manages a logical volume (number of segments) for each logical volume name, and a fourth file 220 that manages the number of slice areas constituting the segment. Have.
The third file 210 has a pointer 211 to the next logical volume, a logical volume name 212, a segment number 213, a pointer 214 to the segment, and a pointer 215 to the tree for facilitating search. , There are as many entries 200 as the number of modules that can communicate with the volume information module M64. The fourth file 220 has a pointer 221 to the next segment, a segment serial number 222, and two slice areas 223 and 224 constituting the segment. The tree has pointers to the slice areas 223 and 224 of the segment of the fourth file 220.
40 and 41 are diagrams showing an example of the management information update function of the volume information module M64.
As shown in FIG. 40, the volume information module M64 includes an update request unit 64d that sends a management information update request to each module.
The update request unit 64 d can send a management information update request to each of the modules 31 to 34 via the communication function unit 29.
41, in addition to the functions of the volume information module M64 shown in FIG. 40, an update completion request unit 64e that requests transmission of management information update completion information to each module and a request to stop updating management information And an update cancel request unit 64f.
Such a function of the volume information module M64 is used when an abnormality is detected from the target module when the management information of a plurality of modules is updated at the same time, or when the volume information module M64 is updating the management information. This is useful when you have to stop updating.
FIG. 42 is a diagram showing an example of a copy request function for data stored in the module, which the volume information module M64 has.
As shown in FIG. 42, the volume information module M64 is stored in the slice area of the module in addition to the update request unit 64d, the update completion request unit 64e, and the update stop request unit 64f shown in FIGS. A copy request unit 64g for instructing data to be copied to a slice area of another module and a copy stop request unit 64h for stopping the copy are provided.
FIG. 43 is a diagram illustrating an example of a function of the volume information module M64 that forcibly terminates processing in each module.
As shown in FIG. 43, in addition to the various functions shown in FIG. 42, the volume information module M64 includes a forced termination command unit 64i that forcibly terminates processing in a specific module or all modules. Accordingly, it is possible to forcibly terminate the processing in the module where the abnormality is detected or to stop the operation of the entire storage system when abnormality is detected in a plurality of modules.
Next, various processing procedures performed by the volume information module M64 will be described.
44 to 46 are diagrams illustrating an example of a processing procedure when the volume information module M64 receives a data storage request.
44 to 46, the storage system includes a module G31, a module H32, a module I33, a module J34, and a volume information module M64. These modules 31 to 34 and 64 have communication function units 25 to 29, respectively. Have.
As shown in FIG. 44, in the storage system of this embodiment, the volume information module M64 receives a data allocation request consisting of a logical volume with four segments from the outside.
As shown in FIG. 45, the volume information module M64 that has received the allocation request allocates unused slice areas as candidates so that slice areas belonging to different modules constitute segments, and the module to which the allocated slice area belongs. To the management information update request.
As shown in FIG. 46, each of the modules 31 to 34 allocates data to the slice area 15 allocated by the volume information module M64 and manages the management information stored in the management information storage unit 18 based on the management information update request. Update information. As a result, the slice area numbers S12 and S13 of the module G31, the slice area number S25 of the module H32, the slice areas numbers S31 and S34 of the module I33, and the slice area numbers S41, S43, S44, and S45 of the module J34 Logical volume L3 is allocated. The logical volume L3 includes segments (S12, S31), (S25, S41), (S13, S44), and (S34, S45).
Thus, since the volume information module M64 assigns the requested logical volume, the assignment rule can be changed relatively easily by changing the program of the volume information module M64.
47 to 49 are diagrams illustrating an example of a processing procedure when the volume information module M64 receives a logical volume release request.
47 to 49, the storage system includes a module G31, a module H32, a module I33, a module J34, and a volume information module M64. The modules 31 to 34 and 64 have communication function units 25 to 29, respectively. Have.
In FIG. 47, the volume information module M64 receives a request to release the logical volume L3 (number of segments 1) from the outside.
As shown in FIG. 48, the volume information module M64 searches for the number of the slice area belonging to the logical volume L3 based on the logical volume information that the logical volume unit 64c has. As a result of the search, since the slice area number S12 of the module G31 and the slice area number S31 of the module I33 belong to L3, a management information update request is transmitted to the modules G31 and I33.
As shown in FIG. 49, the module G31 and the module I33 that have received the management information update request delete the data of the numbers S12 and S31 and update the management information.
Thus, since the logical volume unit 64c of the volume information module M64 has the logical volume information, it is only necessary to send an update request to the necessary modules.
50 to 53 are diagrams illustrating an example of a processing procedure when a predetermined module fails.
50 to 53, the storage system includes a module G31, a module H32, a module I33, a module J34, and a volume information module M64. These modules 31 to 34 and 64 have communication function units 25 to 29, respectively. Have.
In FIG. 50, it is assumed that the module I33 has failed and the volume information module M64 has recognized it.
In FIG. 51, since the volume information module M64 knows that data is stored in the slice areas S32 and S33 from the management information of the module I33, the volume information module M64 secures two slice areas lost due to the failure from other modules. . At that time, the volume information module M64 sorts the unused slice areas so that the segments are constituted by slice areas belonging to different modules. Here, the slice area S12 is selected instead of the slice area S32, and the slice area S41 is selected instead of the slice area S33. Therefore, the volume information module M64 requests the module G31 to secure the slice area S12 and the module J34 to secure the number S41.
In FIG. 52, when the slice area S12 and the slice area S41 are secured, the volume information module M64 transfers the data stored in the slice area S23 to the slice area S23 and the slice area S23 constituting the segment. And a command to copy the data stored in the slice area S14 to the slice area S41 is issued to the slice area S33 and the slice area S14 constituting the segment.
As shown in FIG. 53, when the data copy is completed, the volume information module M64 makes a management information update request to the module G31 and the module J34.
As described above, the volume information module M64 that holds the volume information in a centralized manner performs the restoration processing of the data of the failed module in a unified manner, whereby the data redundancy can be quickly recovered.
54 to 56 are diagrams illustrating an example of a processing procedure when a predetermined module is pulled out.
54 to 56, the storage system includes a module G31, a module H32, a module I33, a module J34, and a volume information module M64. These modules 31 to 34 and 64 have communication function units 25 to 29, respectively. Have.
In FIG. 54, the volume information module M64 receives an extraction request for the module I33 from the outside. Here, a securing request is made to the module G31 and the module J34 in order to secure the slice area S12 and the slice area S41.
As shown in FIG. 55, when the evacuation destination slice area of the data stored in the module I33 is secured, the volume information module M64 copies the data stored in the slice area S33 of the module I33 to S41 of the module J34. And instructing to copy the data stored in the slice area S23 of the module H32 to the slice area S12 of the module G31.
Here, the data stored in S23 of the module H32 is copied to the slice area S12 of the module G31. The volume information module M64 has the number information of the slice areas constituting the segment. This is because the area S23 constitutes a segment and knows the fact that the same data is stored. By taking such a procedure, it is possible to reduce the load bias of the extracted data saving.
As shown in FIG. 56, when the data copy is completed, the volume information module M64 makes a management information update request, determines the related management information, and then pulls out the module I33.
FIG. 57 to FIG. 60 are diagrams illustrating an example of a processing procedure when replacing a module.
The processing procedure illustrated in FIGS. 57 to 60 is a diagram illustrating an example of a processing procedure when module K is added and module J is replaced with module K after module I is extracted.
57 to 60, the storage system includes a module G31, a module H32, a module I33, a module J34, and a volume information module M64. These modules 31 to 34 and 64 have communication function units 25 to 29, respectively. Have.
As shown in FIG. 57, when the volume information module M64 receives an exchange request between the module J34 and the module K35, as shown in FIG. 58, the volume information module M64 makes a request for securing a slice area to the module K35. Here, it is only necessary to secure three slice areas in which data is stored in the module J34. When it is confirmed that three slice areas can be secured, as shown in FIG. 59, the volume information module M64 issues a copy request for the stored data to the module J34.
Here, in order to reduce the load on the entire storage system, copying is performed by sharing from the modules constituting the segment.
As shown in FIG. 60, when the data copying is completed, the volume information module M64 issues an update request for the management information of each of the modules G31, H32, and K35. When the update is completed, the module J34 is pulled out and replaced. finish.
(Fourth embodiment)
Compared to the third embodiment, the fourth embodiment has a volume information module in which the storage system reconstructs the volume information of data distributed and stored in each module, and FIG. 29 of the second embodiment. It differs from the point provided with the access module described in the description. However, since the other points are common, these differences will be described below.
FIG. 61 is a diagram illustrating a configuration example of an information processing system including the storage system according to the fourth embodiment.
In FIG. 61, the storage system 1 receives four modules G31, module H32, module I33, and module J34, each of which has a PC having storage devices 21 to 24, and the corresponding slice when receiving an access request to the storage system 1. Two access modules A41 and B42 each including a PC that directly accesses the area are provided, and each of the four modules 31 to 34 is connected to the network 2 via the two access modules A41 and B42. In addition, the logical volume information of the data distributed and stored in each of the modules 31 to 34 is reconstructed, and a volume information module M64 including a PC that issues a command to each module based on the information is provided. Are connected to the network 2 respectively. Each of the two access modules A41 and B42 receives access from the external computer 3 and can directly access the related modules 31 to 34, and can also access the logical volume information section 64c of the volume information module M64 via the network 2. By accessing, the logical volume information can be referred to.
FIG. 62 is a functional block diagram showing the storage system of the fourth embodiment.
The storage system of this embodiment is the same as that described in the third embodiment, except that it does not have a state management module and has two access modules. The access module is the same as that described in the second embodiment. Accordingly, the same components are denoted by the same reference numerals, and redundant description is omitted.
The storage system shown in FIG. 62 collects management information from modules G31, module H32, module I33, module J34, two access modules A41 and B42, and collects module information and logical volume information. Volume information module M64.
Each of the two access modules A41 and B42 receives an access request from the external computer 3, refers to the module information and logical volume information from the volume information module M64 via the network, and directly sends the access request to the corresponding module. You can intervene. Therefore, the access module does not need to access individual modules to collect information, and only needs to hold a part of the information. Therefore, the load on the communication line for accessing the individual modules can be reduced. it can. In the present embodiment, since the two access modules A41 and B42 have a duplex configuration, even when a load imbalance occurs when accessing the access modules A41 and B42 from the outside, the load imbalance is absorbed. , Safety when accessing from outside can be improved.

  As described above, the storage system of the present invention is useful as a storage device for an information processing system that requires reliability. In particular, even when a PC or server having storage devices with different capacities is combined, Since the usage efficiency of each storage device can be increased, the storage device is suitable for use in an information processing system that is highly likely to be expanded or replaced.

Claims (20)

In storage systems that store data across multiple storage devices,
A storage device having a data storage section having a plurality of sliced slice areas for storing data and a management information storage section for storing management information of the plurality of slice areas;
A storage system comprising a plurality of modules each having a communication function unit having a communication function of data including data stored in the data storage unit and information including management information stored in the management information storage unit. The storage system according to claim 1, wherein the plurality of modules communicate the information with each other by the communication function unit. By storing the same data in each of the slice areas of at least two modules of the plurality of modules, a segment composed of a pair of the slice areas is formed, and redundancy of the data stored in the data storage unit is increased. The storage system according to claim 1, wherein the storage system is secured. The storage system according to claim 1, wherein the management information storage unit stores information relating to a logical volume of sliced areas of a plurality of modules of data for which a storage request has been received. The first module of the plurality of modules has a function of copying data stored in the slice area of the first module to a predetermined slice area of the second module. The storage system according to 1. The management information storage unit of the first module among the plurality of modules receives an update request for management information stored in the management information storage unit from the second module, from the second module. 2. The management information is updated when an update completion request is received, and the management information is restored to a state immediately before the management information update request is received when an update stop request is received. The described storage system. The storage system according to claim 1, wherein each of the plurality of modules includes a forced activation unit that performs forced activation after abnormal termination. The storage system according to claim 1, wherein each of the plurality of modules includes a heartbeat transmission unit that transmits a heartbeat for notifying each operation state. 2. The storage system according to claim 1, further comprising an access module that centrally accepts an access request to the storage system and relays the access request to a corresponding module. The storage system according to claim 9, wherein the access module includes information of data stored in the slice area of each of the plurality of modules, and responds to the access request based on the information. Each of the plurality of modules has a heartbeat sending unit that sends out a heartbeat notifying each operation state,
The storage system according to claim 1, further comprising a state management module that manages an operation state of each of the plurality of modules based on a heartbeat received from the heartbeat transmission unit. 12. The storage system according to claim 11, wherein the state management module includes a forced termination command transmission unit that forcibly terminates the operation of a predetermined module among the plurality of modules. The state management module includes a module state transmission unit for notifying the other modules of the operation states of the plurality of modules recognized by the state management module,
12. Each of the plurality of modules is configured to stop each operation when an operation state received from the state management module is different from an operation state recognized by each module. Storage system. A volume information module that collects the management information from each of the plurality of modules, and reconstructs information related to the logical volume over the slice areas of the plurality of modules of the data that has received a storage request based on the collected management information The storage system according to claim 1, further comprising: 15. The storage system according to claim 14, wherein the volume information module includes an update request unit that makes an update request for the management information to each of the plurality of modules. The volume information module includes an update completion request unit that makes an update completion request for the management information and an update stop request unit that makes an update stop request for the management information for each of the plurality of modules. The storage system according to claim 14. The first module of the plurality of modules has a function of copying data stored in the slice area of the first module to a predetermined slice area of the second module,
15. The storage system according to claim 14, wherein the volume information module designates the predetermined slice area in which data stored in the slice area of the first module is copied to a second module. 15. The storage system according to claim 14, wherein the volume information module includes a forced termination command sending unit that forcibly terminates the operation of each of the plurality of modules. A function of receiving a heartbeat for notifying the operation state of each of the plurality of modules, managing the operation state of each of the plurality of modules, and forcibly terminating the operation of each of the plurality of modules when the heartbeat is stopped. A state management module having
The volume information module receives a notification of an operation state of each of the plurality of modules from the state management module, and instructs the state management module to forcibly terminate a predetermined module among the plurality of modules. The storage system according to claim 14. In addition to accepting access requests to the storage system in a unified manner, the corresponding module is provided with an access module that relays the access requests,
15. The access module, when receiving the access request, acquires information related to the access request from the volume information module and relays the access request based on the acquired information. The described storage system.

Images